Historically, when I've added MIB files to my net-snmp installation, I
do something like this:
(1) mkdir /opt/local/share/snmp/mibs/Cisco
(2) cp ~/somewhere/* /opt/local/share/snmp/mibs/Cisco
(3) echo mibdirs +/opt/vdops/share/snmp/mibs/Cisco >>
/opt/local/etc/snmp/snmp.conf
And, of course, somet
OK, turns out I described the issue incompletely.
By wiping /opt/local/var/snmp/mib_indexes, I can squeeze out a single
successful run of a net-snmp utility:
guru> snmpget -c public localhost sysDescr.0
RFC1213-MIB::sysDescr.0 = STRING: "Linux guru.company.com
2.6.18-308.11.1.el5.centos.plus #1 S
On 8 August 2012 14:09, Stuart Kendrick wrote:
> ==> I don't see a '.index' file anywhere in /opt/local/share/snmp/mibs
> ... is this a problem?
No.
The index file was moved from the directory containing the MIB files,
(which could legitimately be mounted on read-only storage),
to come under the
Hi Dave,
Seems to me that I have read/write to the relevant files ... and that
the contents of the index files are congruent with the contents of the
MIB file directories (although I haven't been precise about this, i.e. I
haven't precisely checked that every file in the MIB file directories
shows
On 8 August 2012 14:46, Stuart Kendrick wrote:
> guru> cat 1 | more
> DIR /opt/local/share/snmp/mibs/Cisco
> CISCO-SYSTEM-MIB CISCO-SYSTEM-MIB
> [...]
> guru> cat 2 | more
> DIR /opt/local/share/snmp/mibs/IETF
> DNS-RESOLVER-MIB DNS-RESOLVER-MIB
> [...]
Can you please post the *whole* of these t
On 8 August 2012 15:56, Stuart Kendrick wrote:
> I have attached these files
OK - I think I've tracked down the problem.
These index files seem to use DOS-style line endings (i.e. CR/LF)
Unfortunately, the code to process them includes some very Unix-centric
thinking, and is explicitly looking f
> OK - I think I've tracked down the problem.
>
> These index files seem to use DOS-style line endings (i.e. CR/LF)
> Unfortunately, the code to process them includes some very Unix-centric
> thinking, and is explicitly looking for the LF character.
> (See 'snmplib/parse.c:add_mibdir()')
>
>I'm
On 9 August 2012 10:44, Stuart Kendrick wrote:
> So now I'm poking through the pre-converted list of MIB files running
> hexdump, looking for end-of-line characters ... and all I'm seeing is
> '0a', which suggests that the MIB files I'm reading were all Unix-style.
>
> Would you identify a particu
On 8/9/2012 3:04 AM, Dave Shield wrote:
> If you look in the IETF folder, you should see five duplicate names
> (four of the form T11-FC*, plus DOCS-IETF-SUBMGT-MIB) In fact, the
> names aren't exact duplicates - there is actually a space at the end
> of the second version. This space is what is c